Exosomes from adipose-derived stem cells alleviate premature ovarian failure via blockage of autophagy and AMPK/mTOR pathway

Establishment of POF mouse model

Female C57BL/6 mice (n = 36, 18–22 g) aged 6-8 weeks were purchased from the experimental animal center of Inner Mongolia People’s Hospital and acclimated to the conditions of a room at a humidity level of 44 ± 2% humidity and 12 h light/12 h dark cycle at 23 ± 3 °C for one week. Food and water were provided ad libitum. Mice with normal estrous cycle were selected by vaginal exfoliated cell smear. The POF mouse model was established by intraperitoneal injection of 50 mg/kg cyclophosphamide (a chemotherapeutic drug) on the first day, and then continuous injections of 8 mg/kg/day in the next 14 days (Yuan et al., 2022). Vaginal exfoliated cell smear shows the disordered estrous cycle in mice, indicating that the POF mouse model is established successfully (Guo et al., 2019). After the model was established, 15 mg/kg of 3-methyladenine (3-MA, an autophagy inhibitor) dissolved in dimethyl sulfoxide (DMSO) was intraperitoneally injected into POF mice twice a week. The wild-type (WT) group consisted of normal, untreated mice. On the 7th and 15th days after 3-MA treatment, all mice were sacrificed by cervical dislocation. Then, ovaries were removed, and their wet weights were recorded. This study was approved by the Institutional Animal Care and Use Committee of Inner Mongolia People’s Hospital (No. 2020-043).

Separation and identification of ADSCs-Exo

Mouse ADSCs (American Type Cell Culture) were routinely cultured in DMEM/F12 medium. Extraction of ADSCs-Exo was performed as previously described with slight modification (Zhou et al., 2022). Cells of 3–5 generations with good growth were cultured in medium without serum for 24 h. When cell confluence reached 80%, cells were washed with PBS thrice and cultured in fresh medium at 37 °C with 5% CO₂ for another 48 h. The cell supernatant was collected after centrifuging for 40 min at 4 °C in an ultrafiltration concentration tube. The exosome-containing concentrate was obtained by centrifuging at 4,000 × g for 40 min and then transferred to a sucrose/heavy water density pad at a concentration of 30%. After sterile membrane filtration and decontamination, ADSCs-Exo were collected from the bottom by centrifuging at 100,000 × g for 120 min at 4 °C. Morphological characteristics of exosomes were observed by transmission electron microscopy (TEM), and the particle-size was analyzed by a Flow Nanoanalyzer (NanoFCM, Nottingham, UK). The protein expression levels of CD63, CD81, TSG101, and HSP70, biomarkers for exosomal surfaces, were assessed through western blot analysis in ADSCs extracted samples, utilizing phosphate-buffered saline (PBS) as the control.

ADSCs-Exo treatment

POF mice received 100 µL of PBS, ADSCs, Exo, Exo + AICAR (0.5g/kg, an AMPK activator), or Exo + Rapa (8 mg/kg, an inducer of mTOR-mediated autophagy) suspension via tail vein injection utilizing a one mL syringe. Mice were injected once every other day from the first day of modeling for 2 weeks. The mice were weighed, and eyeball blood samples were taken on the 7th (n = 3 each group) and 15th (n = 6 each group) day. Mice were sacrificed by cervical dislocation on the 7th and 15th days, and their ovaries were taken out and weighed.

Hematoxylin and eosin (HE) staining

HE staining was conducted by referring to a previously reported method (Wen et al., 2022). Ovarian tissues were collected from mice on the 7th and 15th day after treatment and fixed with 4% paraformaldehyde for 24 h. After gradient alcohol dehydration, transparency, wax dipping, and embedding, tissues were sliced to 5 µm-thick sections. After baking at 60 °C for 2 h, dewaxing with xylene, and hydration with gradient alcohol, sections were stained with hematoxylin aqueous solution for 30 s. Then, sections were stained with eosin for 2 min and then observed under a light microscope (Leica, Wetzlar, Germany).

Enzyme-linked immunosorbent assay (ELISA)

On the 7th and 15th day, eye peripheral blood of mice was taken and centrifuged at 1500 × g at 4 °C for 15 min to obtain the serum supernatant. Levels of serum oxidative stress factors (superoxide dismutase (SOD), malondialdehyde (MDA), and ROS) and serum ovarian function related factors (estradiol (E2) and FSH) in each group were measured by corresponding ELISA kits. Experimental process was operated in strict accordance with the instructions of kits as follows: the SOD kit (SBJ-M0412; Nanjing SenBeiJia, Nanjing, China), the MDA kit (KS13329; Shanghai Keshun Biology, Shanghai, China), the ROS kit (SBJ-M0608; Nanjing SenBeiJia), the E2 kit (E-EL-0150l, Elabscience, Wuhan, China), and the FSH detection kit (E-EL-M0511c; Elabscience).

Immunofluorescence

Ovarian tissues were fixed with 4% paraformaldehyde for 24 h at room temperature. Tissue sections underwent microwave-assisted antigen retrieval at 92–96 °C for 15 min, followed by blocking with 5% BSA at 37 °C for 1 h. The sections were then incubated with FSHR antibody (1:200; AF5242; Affinity, USA) overnight at 4 °C, followed by a 1 h incubation at 37 °C with Goat Anti-Rabbit IgG H&L/AF488 secondary antibody (1:200; bs-0295G-AF488; Bioss, China). Finally, the sections were counterstained with DAPI and visualized under a fluorescence microscope (BX53; Olympus, Tokyo, Japan).

TUNEL assay

Ovarian tissues were collected from mice on the 15th day for TUNEL detection (Luo et al., 2022). Ovarian tissue sections in each group were routinely dewaxed and hydrated, followed by the addition of 50 µL proteinase K working solution for digesting at 37 °C for 30 min. Sections were incubated with 5 µL terminal deoxynucleotidyl transferase (TdT) enzyme, 45 µL fluorescence labeling solution, and 50 µL TUNEL detection solution (Beyotime, Jiangsu, China) at 37 °C for 60 min without light. After rinsed with PBS for three times, 4′,6-diamidino-2-phenylindole (DAPI) dye was dropped on sections and incubated for 10 min at ambient temperature. Sections were sealed with anti-fluorescence quenching sealing solution and then photographed under a fluorescence microscope (Olympus, Japan).

Transmission electron microscopy (TEM)

Ovarian tissues were collected from mice on the 15th day after treatment and subjected to TEM (Dou et al., 2022). Tissues were cut into 1 µm sections and fixed in 2.5% precooled glutaraldehyde solution at 4 °C overnight. After rinsed with PBS thrice, the ovarian tissues were stained with 1% osmic acid, dehydrated with gradient ethanol and 90% acetone, and embedded in ultra-thin sections. After toluidine blue staining, the ovarian tissues were examined under a light microscope (Olympus, Japan). The ultrastructure of ovarian GCs and the changes in mitochondrial autophagy were photographed by TEM (Shanghai Weihan Photoelectric Technology, Shanghai, China).

Cell culture and treatment

Human ovarian granulosa-like tumor (KGN) cell line (Feiya Biotechnology Co., Ltd., Haian, China) were used for in vitro study. KGN cells were cultured in Roswell Park Memorial Institute (RPMI) 1640 Medium (Thermo Fisher Scientific, Waltham, MA, USA) with 10% fetal bovine serum (Gibco) in a humidified incubator at 37 °C with 5% CO₂. Cells were divided into four groups: (1) control group (no treatment); (2) cyclophosphamide (CTX) group, in which cells were treated with 250 µM CTX for 48 h to establish POF cell model (Liu et al., 2022); (3) CTX + ADSCs-Exo group, in which cells were co-treated with 250 µM CTX and 10 µg/mL ADSCs-Exo for 48 h; (4) CTX + ADSCs-Exo + Rapa (rapamycin) group, in which cells were pre-treated with 5 µM rapamycin for 1 h and then co-cultured with 250 µM CTX and 10 µg/mL ADSCs-Exo for 48 h.

Cell counting kit (CCK)-8 assay

KGN cell viability was assessed using CCK-8 assay (Yeasen, Shanghai, China). Cells were seeded into a 96-well plate (2 × 10³ cells/well). Next day, CCK-8 solution (10 µL) was added to each well for incubation for 48 h. The OD_{450 nm} values were measured using a microplate reader (Bio-Tek, Wuxi City, China).

Flow cytometry

KGN cell apoptosis was detected as previously described (Lin et al., 2021). Cells were cultured in six-well plates (5 × 10⁵ cells/well) overnight. Then, cells were collected and resuspended in binding buffer, followed by incubating with Annexin V-FITC and PI (Beyotime, Jiangsu, China) for 15 min in the dark. Subsequently, cell apoptosis was detected by flow cytometry (BD FACSCalibur).

Western blotting

Western blotting was performed as described elsewhere (Guo, Zhu & Sun, 2020). Total protein was extracted from ovarian tissues and KGN cells using radioimmunoprecipitation (RIPA) lysis solution containing 1% phenylmethyl sulfonyl fluoride (PMSF) and then was quantified with a bicinchoninic acid (BCA) protein concentration kit (Beyotime, China). Proteins were separated by 10% sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred to polyvinylidene fluoride (PVDF) membrane (Beyotime, Jiangsu, China) via electrophoresis under the condition of 200 mA constant current for 90 min. After blocking with 10% skim milk (Beyotime, Jiangsu, China) for 2 h, membranes were incubated with primary antibodies overnight at 4 °C. Then, membranes were incubated with goat anti-mouse IgG secondary antibody (1:2,000, Abcam, Cambrdige, UK) at room temperature for 2 h. Proteins were developed with ECL chemiluminescence solution and exposed in Alpha InnotechFluorchem SP fluorescence chemiluminescence gel image analysis system (Invitrogen, Waltham, MA, USA). Relative protein expression was calculated by normalizing to GAPDH. The primary antibodies used in this study were the AMPK antibody (1:1,000, ab32047, Abcam, Cambridge, UK), p-AMPK antibody (1:1,000, ab32047, Abcam), mTOR antibody (1:1,000, CSB-PA208208; Cusabio, Hubei, China), p-mTOR antibody (1:1,000, CSB-PA271384; Cusabio), Beclin-1 antibody (1:1,000, ab210498; Abcam), LC3II/LC3I antibody (1:1,000, 12741T; CST), Bcl-2 antibody (1:1,000, ab182858; Abcam), CD63 antibody (1:1,000, ab217345; Abcam), CD81 antibody (1:1,000, DF2306; Affinity, Cincinnati, OH, USA), TSG101 antibody (1:1,000, DF8427; Affinity), HSP70 antibody (1:1,000, AF5466, Affinity) and the GAPDH antibody (1:1,000, ab245355; Abcam).

Statistical analysis

All data were presented in the form of mean ± standard deviation. Animal experiments had six biological and three technical replicates; cell experiments had three of each. Differences among groups were compared by one-way ANOVA, followed by Tukey’s test. All statistical analysis was completed on the GraphPad 7.0 (GraphPad Software, La Jolla, CA, USA). P < 0.05 was considered to indicate a statistically significant difference between groups.

Inhibition of autophagy promoted ovarian growth and functional recovery, and suppressed oxidative stress in POF mice

In order to evaluate the effect of autophagy on ovarian function in POF, model mice were induced by cyclophosphamide and treated with autophagy inhibitor 3-MA (Fig. 1). Compared to WT mice, ovarian tissues of POF mice were smaller and had lower weight. Treatment of 3-MA for 15 days recovered the ovarian growth of POF mice (p < 0.01, Fig. 2A). Subsequently, the effects of autophagy inhibition on pathological tissue injury in POF mice were evaluated. HE staining showed a decrease of sinus follicles, GC layer, and corpus luteum, and an increase of atretic follicles in the POF group compared to the WT group. These pathologies were alleviated by 3-MA treatment in POF mice. In addition, the pathological injury was more obvious on the 15th days than on the 7th days after 3-MA treatment (Fig. 2B). On the other hand, ELISA was used to evaluate the levels of ovarian function related factors E2 and FSH. In POF, follicles, estrogen, and progesterone dramatically decrease in ovaries, resulting in an increased serum FSH level and a decreased E2 level (Fu et al., 2017). Our results showed that the levels of E2 and FSH in the POF group were significantly decreased and increased compared with that in the WT group, respectively (p < 0.01). 3-MA intervention for 15 days markedly enhanced E2 level and down-regulated FSH level in POF mice (p < 0.05, Fig. 2C). Furthermore, MDA, ROS, and SOD are critical biomarkers of oxidative stress. ELISA showed that on the 15th day, the contents of SOD in the POF group were substantially decreased, whereas ROS and MDA levels were prominently increased compared with that in the WT group (p < 0.01). 3-MA intervention enhanced SOD, and declined ROS and MDA levels in POF mice on the 15th day (p < 0.05, Fig. 2C).

Inhibition of autophagy alleviated GC apoptosis in POF mice

Apoptosis of ovarian GCs is the initiating factor of POF (Liu et al., 2021). We found that the expression level of FSHR, a known GC biomarker, was significantly reduced in the POF group compared to the WT group (p < 0.01). Treatment with 3-MA successfully upregulated the FSHR level in the POF group, as evidenced by the results (p < 0.01, Fig. S1). TUNEL showed that the apoptosis rate of GCs in the POF group was higher than that in the WT group, which was decreased by 3-MA treatment (Fig. 3A). In addition, TEM was used to observe the ultrastructure of ovarian GCs and the changes of mitochondrial autophagy. Results showed that compared with the WT group, the nuclei of oocytes changed from prismatic to vacuolar in the POF group. Also, the number of mitochondria in GCs decreased, while the number of autophagosomes increased in the POF group. Administration of 3-MA recovered the morphology of GCs and reduced the number of autophagosomes (Fig. 3B). Bcl-2 and Beclin-1 are key regulators of autophagy (Hill, Wrobel & Rubinsztein, 2019; Xu et al., 2020), and LC3II/LC3I can target autophagosomes to mitochondria and induce mitochondrial autophagy (Unal et al., 2021). Western blotting showed that, compared with the WT group, Beclin-1 and LC3II/LC3I levels were significantly increased in the POF group, while Bcl-2 was decreased (p < 0.01). 3-MA treatment markedly downregulated the protein levels of Beclin-1 and LC3II/LC3I and upregulated Bcl-2 level in POF mice (p < 0.05, Fig. 3C).

Inhibition of autophagy reduced GC apoptosis by retarding the AMPK/mTOR pathway in POF mice

Studies have confirmed that mTOR can act as an “activity switch” in the regulation of autophagy (Kazibwe et al., 2020; Wang & Zhang, 2019). The AMPK/mTOR pathway plays an important regulatory role in the process of autophagy (Li et al., 2019). In addition, the AMPK/mTOR pathway implicates oocyte senescence and female reproduction (Wang et al., 2021a). We further explored the role of AMPK/mTOR pathway during 3-MA treatment for POF mice. Western blotting showed that p-AMPK/AMPK expression was increased, and p-mTOR/mTOR level was significantly decreased in POF mice compared with that in WT mice, while reversed by 3-MA treatment (p < 0.05, Fig. 3C).

ADSCs-Exo ameliorated chemotherapy-induced pathological damage in POF mice

Subsequently, we evaluated the effect of ADSCs-Exo in the treatment of POF. In the present study, exosomes isolated from ADSCs exhibited a circular morphology under TEM, with the average particle-size of 56.67 nm (Fig. 4A). Western blotting showed that the surface specific marker proteins CD63, CD81, TSG101, and HSP70 were expressed in ADSCs-Exo compared with that in the control group (Fig. 4B). Firstly, the morphology of ovarian tissue was observed after ADSCs-Exo treatment. The results showed that on the 15th day, ovarian size and weight in the ADSCs group and Exo group were substantially increased compared with the POF group (p < 0.05, Fig. 4C). Furthermore, HE staining showed that the number of follicles in the ADSCs group and Exo group was increased, while atresia follicles were decreased (Fig. 4D).

ADSCs-Exo inhibited GC apoptosis and autophagy by impeding the AMPK/mTOR pathway

Then, ELISA showed that on the 15th day, compared with the POF group, ADSCs and Exo markedly increased E2 level and decreased the contents of FSH (p < 0.05, Fig. 5A). Likewise, ADSCs and Exo markedly increased SOD level and decreased the contents of ROS and MDA in POF mice (p < 0.05, Fig. 5A). On the other hand, TUNEL showed that, ADSCs-Exo treatment reduced the apoptosis rate of GCs in POF mice (Fig. 5B). Subsequently, western blotting presented that, compared with the POF group, the protein levels of Beclin-1 and LC3II/LC3I in ADSCs and Exo group were dramatically downregulated, and Bcl-2 level was upregulated. Finally, the role of the AMPK/mTOR pathway in ADSCs-Exo treatment for POF mice was verified. Western blotting showed that p-AMPK/AMPK level was decreased, and p-mTOR/mTOR was significantly increased in ADSCs and Exo groups compared with that in the POF group (p < 0.05, Figs. 5C and 5D). Administering AICAR, an AMPK activator, reversed the inhibitory effect that ADSCs-Exo had on POF (p < 0.05, Fig. 6). Furthermore, administering Rapa—an inducer of mTOR-mediated autophagy—reversed the protective effect exerted by ADSCs-Exo against POF (p < 0.05, Figs. 7A and 7B).

Further, CTX-treated KGN cells were treated with ADSCs-Exo or/and Rapa to confirm the mechanism of ADSCs-Exo inhibiting GC autophagy in POF. CTX (an alkylating agent) has a pro-apoptotic adverse effect on ovarian GCs (Li et al., 2021b), which was used to treat human ovarian granulosa-like tumor (KGN) cells to construct the cell model of POF. Using the CCK-8 assay, we determined that the IC50 value for ADSCs-Exo in treating CTX-treated KGN cells was 10 µg/Ml (Fig. 8A). Consequently, a concentration of 10 µg/mL for ADSCs-Exo was utilized in subsequent experiments involving the treatment of CTX-treated KGN cells. CCK-8 and flow cytometry showed that CTX inhibited the viability and promotes the apoptosis of KGN cells (p < 0.01). ADSCs-Exo treatment alleviated CTX-induced apoptosis, whereas the addition of rapamycin repressed the effect of ADSCs-Exo on CTX-treated KGN cells (p < 0.05, Figs. 8B and 8C). Regarding the mechanism, CTX upregulated the levels of Beclin-1 and LC3II/LC3I, and downregulated the Bcl-2 level in KGN cells (p < 0.01). ADSCs-Exo administration reduced CTX-induced autophagy, evidenced by the decreased Beclin-1 and LC3II/LC3I as well as the increased Bcl-2 level (p < 0.01). The addition of rapamycin weakened the inhibitory effect of ADSCs-Exo on autophagy in CTX-treated KGN cells (p < 0.05, Fig. 8D). In addition, expression of p-AMPK was upregulated, and that of p-mTOR was downregulated in CTX-treated KGN cells (p < 0.01). ADSCs-Exo treatment decreased the p-AMPK/AMPK level and increased the p-mTOR/mTOR level in CTX-treated KGN cells, whereas rapamycin addition reversed the effect of ADSCs-Exo (p < 0.05, Fig. 8D).